multi-variate distribution

[Paolo Carlini]

On 09/05/2012 01:57 AM, Ulrich Drepper wrote:
> On Mon, Sep 3, 2012 at 1:41 PM, Ulrich Drepper <drepper@gmail.com> wrote:
>> I agree with these points and changed the code.  The overall question
>> stands, are these the best interfaces for the job?
> No comments?
In general I agree with Daniel that additional distributions are great 
(assuming we do the work in ext/, as already discussed). One minor 
detail would be that in the past, with Gaby too, we agreed that code in 
ext should stil use "uglified" names, in principle. Thus, eg, 
__beta_distribution. Do you see a problem with that? Another minor 
detail, I think it would be nice to have a minimum of testcases together 
with the code, just basing things like explicit instantiations, check 
the typedefs, a minimum of generation, just to be sure that nothing is 
trivially broken for some reason.

Thanks,
Paolo.